標題: 魚類品質光學即時標準化檢測系統開發研究
Development of Normalized Real Time Optical Detection System for Fish Quality
作者: 蕭子健
Hsiao Tzu-Chien
國立交通大學資訊工程學系(所)
關鍵字: 非侵入式檢測;螢光光譜學;多變數演算法
公開日期: 2010
摘要: 目前台灣在漁產品的鮮度判別上,大多採用官能檢測法進行檢查,若逢疑慮時 便對其進行侵入性與破壞性檢測法來進行檢驗。而目前的檢測法與其判定方式在各 國法規標準不一的條件下未能有效比對,故許多國家及研究團隊皆開始投入非侵入 性魚體鮮度法則之研究。本計畫採用生物體組之與光學互動之特性,亦即是光學取 樣-非侵入式量測原理,架構“即時魚類矩陣螢光光譜鮮度檢測儀”,包含光纖系 統、光譜儀以及多波段光源,將以多重波段光源個別照射魚體,擷取獨立之光譜資 訊,建立矩陣光譜圖。然後採用多維矩陣光譜多變數演算法與傳統檢測法的結果進 行比對、驗證以及資料庫訓練,以達到建立“魚類品質光學即時標準化檢測系統”研 究開發之目的。希望此系統完成後,不僅可以建構不同魚種的各部位標準組織光譜 資料庫、特異性多變數分析演算法、嚴謹的比對標準外,亦可減少漁民因各國不同 法規所造成的漁獲損失,亦足以提升台灣漁產品競爭力、打通國際市場脈絡,並對 其國際上的品牌形象有長足幫助。
At present sensory methods are the most accurate and widely used methods to evaluate fish freshness. When fish products have questions through sensory method, they will be checked by invasive examines. Now the standard rules are different and not objective at each place so noninvasive standard method for fish freshness are developed by many country. The goal of our proposal is setup as “normalized real time optical detection system for fish quality “.We will acquire fluorescence matrix spectra by our real time fish freshness detection optic system, and then build a new detected model by n way matrix spectra multivariable analysis to compare with common freshness rules. We will hope the results of our research will make Taiwan fish products boost for more national competitions.
官方說明文件#: NSC99-2221-E009-124
URI: http://hdl.handle.net/11536/100756
https://www.grb.gov.tw/search/planDetail?id=2121443&docId=339639
顯示於類別:研究計畫